Wednesday, December 4, 2024

Looking at my original 2024 forecast and making a new one for 2028

 So....back before I had polling, i tried to predict the election using the data I had from the 2020 election and applying an offset to the data. Basically, public opinion shifted 6 points from 2020 even as early as 2023, so I took 2020's results, and shifted them 6 points. And...actually? I kind of nailed it. 

I actually got the right result, 226-312, although that's not too out of the ordinary. I mean 226-312 was a very wide target to hit, given the gulf between the swing states and the next several states was a gap that was around 3 points wide. It was actually one of the most likely possible outcomes in the entire spectrum of possible outcomes. All you needed was a red wave to swing the seven swing states, but not one so large it started digging into the next row of states like virginia, minnesota, or new hampshire, and yeah, that's where the results ended up. Look, I even predicted some movement in NJ toward being a swing state based on that data, although I couldnt predict how extreme the shift would be. 

And that's the caveat. I was off in this initial prediction. Most swing states were within 3-4 points, and here I predicted that they would be between 3 and 7 points. So...all in all...I was off by maybe 3 from the actual results. But again, that's how wide that 226-312 target was. But yeah, the results did follow the overall trend here. I was off by a little overall, but still, if you shifted that map back to the left about 3 points, you get something similar to what actually happened, give or take a bit. 

So with that said, what can we predict for 2028? Well, I'm going to be honest. We don't know. Public opinion is going to shift, we don't know in which direction and how much, but common sense and conventional wisdom would suggest, that given Trump just won and has all three branches of government, that unless he's successful and popular, the backlash will shift 2028 to the left. By how much? We don't know. 2016 to 2020 was actually about 2 points to the left. 2020 to 2024 was about 3-4 back to the right. So we're probably looking at anywhere from 2-4 points back to the left if I had to guess 2028. it depends on a lot of variables. What happens in trump's term, how badly does he bungle things? Do we still have democracy? Are the voters rational enough to understand trump F-ed up when he F-s up, etc. We don't know. 

So....for now, let's assume a 3 point shift to the left, given I'd anticipate a shift somewhere between 2 and 4 points left. Btw, we could see as little as no shift left or even a gain, but we could also see voters reject trump en masse with the shift being much more pronounced to the left. 

So let's assume 3 as a baseline, and let my model do the talking, what does the future 2028 map look like?

I would say, it should look something like this:

So, assuming a 3 point shift, this is what happens to the map. Maine, New Mexico, New Jersey, Virginia, and NE2 all go back to functionally being safe states, similar to 2020 for the most part. New Jersey shifted more than the average so it might be a bit closer than it was in 2020. Minnesota is a D+7 state, similar to 2020. New hampshire also goes back in the likely D column.

The rust belt states of Michigan, Wisconsin, and Pennsylvania all become slightly lean D. Pennsylvania remains the tipping point, because this is a model based on 2024's results after all, and we have a 63% chance of winning, which is almost exactly what I would have given Biden in 2020 had I not weighted polls. 

Georgia is more tilt D. I suspect that one could become lean D in practice, Georgia is the one state that seemed to shift left in the 2024 election cycle, with Atlanta being the one metropolis that actually seemed to shift left in a sea of the rest of the country shifting right. 

Nevada and North Carolina are tilt R. I kind of expect NC to remain stubbornly republican, although Nevada could shift back to the blue column. 

Arizona, I think that's gonna remain lean R. It was the reddest swing state almost consistently and democrats underperformed there.

ME2 is a likely R at R+7, I don't expect it to shift left. Ohio is a borderline likely/safe R. 

Given the margins of Florida, Alaska, Iowa, and Texas, I'd take them out of the swing state column. Unless they swing back with the above average ferocity by which they swung right this election cycle, it's gonna take a miracle to dislodge them from the GOP. At this point, they're as hard for democrats to win as states like Illinois, Colorado, and New York are for republicans. I'm not sure whether that should terrify democrats about the future prospects of those normally safe blue states, or really hammer home just how unwinnable Florida, Texas, an Iowa are for the foreseeable future. We're screwed with them. We can't win them. Like for all the demographics is destiny talk, no, that aspect of our coalition imploded this election cycle. 

And yeah. That's...where I would place 2028 right now, knowing NOTHING about how the next 4 years will play out. 2028 is shaping up to possibly be a 2020 repeat, with narrower margins. I suspect that Vance probably won't be able to defeat the democratic challenger, and the democrats could probably squeak by, even with a neoliberal. Like, this is just a model assuming politics as usual and natural backlash against the GOP in power deenergizing the republican base and energizing the democratic base. It doesnt account for anything regarding the individual candidates, how popular they are, what events shape the next election. And things could shift in either direction. We could have the country really like trump and vance remains a popular successor, where the dems put up some crappy candidate who can't win, in which case we might get a R/D+0 outcome or thereabouts where the dems are in for another butt whooping (see 2004, 1996, 1984 for examples of this kind of phenomenon in practice), or we might see the country turn hard on the republicans, as Vance or whoever replaces Trump lacks charisma and we get a once in a generation realigning candidate come out of nowhere (in which case we could see a D+10 outcome or something relative to now for all i know....think 2008 or 1992 for a relatively recent example of THAT). We don't know. That's the thing. There's uncertainty. Thankfully my model is probabilistic, and with a 4 point MOE meaning an 8 point shift in either direction with a 95% confidence interval, we're talking a 95% chance of the result being between R+5 and D+11, with D+3 being the median outcome. That actually seems reasonable. I actually think more realistically, we are gonna have something like D+2-4. So D+3 is a nice baseline.

And that gives the democrats a 5 in 8 chance of winning the next election. I think that's...actually pretty reasonable. Of all possible outcomes, 37% have the republicans winning, and 63% have the democrats winning. If we assume a 3 point shift left, that's actually perfectly reasonable. Democrats win by about 1.3, which isn't a huge win, but it isn't a loss either. And yeah. Given 2016, 2020, and 2024, I think that's actually a pretty reasonable outcome because things are so polarized you only see things shift like 2-4 points election cycle to election cycle. So yeah. That's where, if I had to guess, 2028 is going to go.

No comments:

Post a Comment